INDEX

Explanations

phrases related to discovering a person in a specific location

oai_token-act-pair · gpt-3.5-turbo

references to individuals being found in a deceased or incapacitated state

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

invoke

-0.86

 Flavoring

-0.77

ãĤ¡

-0.71

/+

-0.67

 Flavor

-0.66

 sponsoring

-0.65

 specialization

-0.64

 ambassadors

-0.63

 Blazing

-0.62

 nick

-0.62

POSITIVE LOGITS

 slumped

1.12

 lifeless

1.11

 unconscious

1.03

responsive

1.03

 stret

0.98

 charred

0.98

 handcuffed

0.98

 beaten

0.94

rawled

0.92

 bleeding

0.92

Activations Density 0.296%

phrases related to discovering a person in a specific location

references to individuals being found in a deceased or incapacitated state

No Comments

No Known Activations